ds2_midterm
Ruihan Zhang
2023-03-28
##
## Call:
## lm(formula = .outcome ~ ., data = dat)
##
## Residuals:
## Min 1Q Median 3Q Max
## -67.936 -13.844 -1.397 10.958 214.844
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -3.525e+03 1.866e+02 -18.894 < 2e-16 ***
## age -8.418e-02 1.760e-01 -0.478 0.63260
## gender -6.721e+00 1.386e+00 -4.849 1.38e-06 ***
## race2 -4.587e+00 3.202e+00 -1.433 0.15214
## race3 -2.286e+00 1.751e+00 -1.306 0.19189
## race4 3.600e-03 2.367e+00 0.002 0.99879
## smoking1 4.008e+00 1.543e+00 2.597 0.00949 **
## smoking2 7.161e+00 2.503e+00 2.861 0.00429 **
## height 2.060e+01 1.100e+00 18.723 < 2e-16 ***
## weight -2.239e+01 1.162e+00 -19.275 < 2e-16 ***
## bmi 6.677e+01 3.303e+00 20.212 < 2e-16 ***
## hypertension 1.838e+00 2.307e+00 0.797 0.42565
## diabetes 1.562e+00 1.947e+00 0.803 0.42232
## SBP 1.048e-01 1.491e-01 0.703 0.48218
## LDL -1.018e-02 3.601e-02 -0.283 0.77741
## vaccine -8.737e+00 1.415e+00 -6.175 8.67e-10 ***
## severity 9.505e+00 2.314e+00 4.108 4.22e-05 ***
## studyB 4.580e+00 1.739e+00 2.634 0.00854 **
## studyC 3.436e-01 2.183e+00 0.157 0.87492
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 25.77 on 1383 degrees of freedom
## Multiple R-squared: 0.3552, Adjusted R-squared: 0.3468
## F-statistic: 42.32 on 18 and 1383 DF, p-value: < 2.2e-16
## [1] 20.59645## 19 x 1 sparse Matrix of class "dgCMatrix"
## s1
## (Intercept) -93.205765308
## age .
## gender -5.919136420
## race2 -0.010213828
## race3 -1.777671736
## race4 .
## smoking1 1.948592247
## smoking2 6.302942295
## height 0.247367973
## weight -0.890672253
## bmi 5.573001483
## hypertension 0.275670589
## diabetes 1.355426214
## SBP 0.105877149
## LDL -0.001277377
## vaccine -8.352237305
## severity 10.552981481
## studyB 5.646347276
## studyC .
## [1] 24.77985
## [1] 0## alpha lambda
## 1001 1 0.1353353
## [1] 20.05926## Data: X dimension: 1402 18
## Y dimension: 1402 1
## Fit method: oscorespls
## Number of components considered: 12
## TRAINING: % variance explained
## 1 comps 2 comps 3 comps 4 comps 5 comps 6 comps 7 comps
## X 9.148 17.36 26.52 32.78 38.19 44.15 47.09
## .outcome 16.443 17.59 17.79 18.03 18.58 20.31 27.87
## 8 comps 9 comps 10 comps 11 comps 12 comps
## X 50.86 54.19 57.48 62.98 68.02
## .outcome 32.53 34.50 35.47 35.52 35.52
## [1] 20.59837| nprune | degree | |
|---|---|---|
| 14 | 15 | 1 |
## (Intercept) h(bmi-31.7) h(31.7-bmi) h(bmi-35.1) vaccine
## -170.250124 -34.348898 37.514010 58.186555 -8.594937
## h(height-159.3) h(159.3-height) h(90.6-weight) h(bmi-29.4) gender
## -16.693010 17.827204 -4.702130 12.887379 -5.747563
## severity h(bmi-22.8) studyB h(height-156.1) smoking2
## 9.216420 23.362400 4.580239 12.527257 6.582499
## [1] 19.54443##
## Family: gaussian
## Link function: identity
##
## Formula:
## .outcome ~ gender + race2 + race3 + race4 + smoking1 + smoking2 +
## hypertension + diabetes + vaccine + severity + studyB + studyC +
## s(age) + s(SBP) + s(LDL) + s(bmi) + s(height) + s(weight)
##
## Parametric coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 45.7757 1.8958 24.146 < 2e-16 ***
## gender -5.8950 1.2196 -4.833 1.49e-06 ***
## race2 -4.7692 2.8264 -1.687 0.09176 .
## race3 -2.3028 1.5440 -1.491 0.13606
## race4 -2.2816 2.0921 -1.091 0.27565
## smoking1 3.5670 1.3643 2.614 0.00903 **
## smoking2 7.5238 2.2095 3.405 0.00068 ***
## hypertension 2.0701 1.2218 1.694 0.09044 .
## diabetes 2.1927 1.7190 1.276 0.20233
## vaccine -8.1553 1.2483 -6.533 9.06e-11 ***
## severity 8.8627 2.0398 4.345 1.50e-05 ***
## studyB 4.3530 1.5327 2.840 0.00458 **
## studyC 0.1153 1.9233 0.060 0.95222
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Approximate significance of smooth terms:
## edf Ref.df F p-value
## s(age) 1.212e-09 9 0.000 0.971
## s(SBP) 2.143e-09 9 0.000 0.976
## s(LDL) 2.499e-09 9 0.000 0.873
## s(bmi) 8.753e+00 9 79.687 < 2e-16 ***
## s(height) 7.240e+00 9 6.053 < 2e-16 ***
## s(weight) 4.320e+00 9 5.018 5.48e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## R-sq.(adj) = 0.497 Deviance explained = 50.9%
## GCV = 523.55 Scale est. = 511.11 n = 1402
## NULL
## [1] 19.36985##
## Family: gaussian
## Link function: identity
##
## Formula:
## .outcome ~ gender + race2 + race3 + race4 + smoking1 + smoking2 +
## hypertension + diabetes + vaccine + severity + studyB + studyC +
## s(age) + s(SBP) + s(LDL) + s(bmi) + s(height) + s(weight)
##
## Parametric coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 45.7757 1.8958 24.146 < 2e-16 ***
## gender -5.8950 1.2196 -4.833 1.49e-06 ***
## race2 -4.7692 2.8264 -1.687 0.09176 .
## race3 -2.3028 1.5440 -1.491 0.13606
## race4 -2.2816 2.0921 -1.091 0.27565
## smoking1 3.5670 1.3643 2.614 0.00903 **
## smoking2 7.5238 2.2095 3.405 0.00068 ***
## hypertension 2.0701 1.2218 1.694 0.09044 .
## diabetes 2.1927 1.7190 1.276 0.20233
## vaccine -8.1553 1.2483 -6.533 9.06e-11 ***
## severity 8.8627 2.0398 4.345 1.50e-05 ***
## studyB 4.3530 1.5327 2.840 0.00458 **
## studyC 0.1153 1.9233 0.060 0.95222
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Approximate significance of smooth terms:
## edf Ref.df F p-value
## s(age) 1.212e-09 9 0.000 0.971
## s(SBP) 2.143e-09 9 0.000 0.976
## s(LDL) 2.499e-09 9 0.000 0.873
## s(bmi) 8.753e+00 9 79.687 < 2e-16 ***
## s(height) 7.240e+00 9 6.053 < 2e-16 ***
## s(weight) 4.320e+00 9 5.018 5.48e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## R-sq.(adj) = 0.497 Deviance explained = 50.9%
## GCV = 523.55 Scale est. = 511.11 n = 1402
## NULL
## [1] 19.36985##
## Call:
## summary.resamples(object = res)
##
## Models: lm, lasso, elastic_net, pls, gam1, gam2, mars
## Number of resamples: 50
##
## MAE
## Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
## lm 14.92169 16.79714 17.53143 17.48965 18.20362 20.65141 0
## lasso 14.44659 16.68324 17.68983 17.84015 18.87016 21.82297 0
## elastic_net 14.46166 16.27802 17.01237 17.06807 17.87792 20.75223 0
## pls 14.92145 16.79669 17.53011 17.48984 18.20224 20.65250 0
## gam1 13.62327 15.14377 15.84965 16.05984 16.72174 19.57934 0
## gam2 13.07693 15.16505 15.93548 16.05069 17.01551 18.28793 0
## mars 13.94882 15.45564 16.18787 16.17158 16.67532 19.10351 0
##
## RMSE
## Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
## lm 19.59117 23.05018 25.24154 25.85722 28.59408 40.90569 0
## lasso 18.42737 23.38819 27.45523 28.39035 32.31464 48.24480 0
## elastic_net 18.40971 22.26193 25.41912 26.17550 29.58809 44.03522 0
## pls 19.58634 23.05011 25.24227 25.85717 28.59370 40.90569 0
## gam1 18.83947 21.49673 23.12193 23.89618 26.02227 36.39843 0
## gam2 17.03498 21.44737 23.28293 23.82334 26.00317 32.92817 0
## mars 18.84077 22.41804 23.24004 24.15887 26.00755 34.08055 0
##
## Rsquared
## Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
## lm 0.119520680 0.2365440 0.3266709 0.3285216 0.4085749 0.5970516 0
## lasso 0.003666023 0.1352594 0.1871645 0.1856959 0.2323649 0.3296231 0
## elastic_net 0.102361853 0.2560678 0.3266106 0.3125386 0.3777359 0.5246915 0
## pls 0.119510263 0.2365178 0.3266308 0.3285227 0.4085783 0.5970333 0
## gam1 0.108667896 0.2926807 0.3566997 0.4136214 0.5908192 0.7489117 0
## gam2 0.197959913 0.3015594 0.3958174 0.4303463 0.5515767 0.7493408 0
## mars 0.128490431 0.2906790 0.3755500 0.4190045 0.5525225 0.7427045 0
## [1] 23.89618
## [1] 19.36985##
## Family: gaussian
## Link function: identity
##
## Formula:
## recovery_time ~ gender + race + +smoking + hypertension + diabetes +
## vaccine + severity + study + s(age) + s(SBP) + s(LDL) + s(bmi) +
## s(height) + s(weight)
##
## Parametric coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 45.8507 2.0477 22.391 < 2e-16 ***
## gender -5.8254 1.2283 -4.743 2.33e-06 ***
## race2 -4.7591 2.8307 -1.681 0.092944 .
## race3 -2.3053 1.5474 -1.490 0.136504
## race4 -2.3204 2.0956 -1.107 0.268361
## smoking1 3.5753 1.3667 2.616 0.008992 **
## smoking2 7.5368 2.2135 3.405 0.000681 ***
## hypertension 1.7696 2.0366 0.869 0.385059
## diabetes 2.1977 1.7229 1.276 0.202329
## vaccine -8.1433 1.2508 -6.510 1.05e-10 ***
## severity 8.9118 2.0454 4.357 1.42e-05 ***
## studyB 4.3804 1.5364 2.851 0.004421 **
## studyC 0.1369 1.9278 0.071 0.943408
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Approximate significance of smooth terms:
## edf Ref.df F p-value
## s(age) 1.000 1.000 0.115 0.735
## s(SBP) 1.000 1.000 0.058 0.810
## s(LDL) 1.000 1.000 0.018 0.893
## s(bmi) 8.738 8.973 78.931 < 2e-16 ***
## s(height) 7.179 8.173 5.161 2.08e-06 ***
## s(weight) 5.084 6.224 5.775 5.00e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## R-sq.(adj) = 0.496 Deviance explained = 50.9%
## GCV = 526.4 Scale est. = 512.51 n = 1402
